The Hippocampus in Pigeons Contributes to the Model-Based Valuation and the Relationship between Temporal Context...

Area of Science:

• Neuroscience
• Comparative Cognition
• Behavioral Economics

Background:

• Model-based decision-making relies on representing relationships between states.
• The mammalian hippocampus is crucial for learning these relationships.
• Neural mechanisms of model-based learning in birds' hippocampi are under-explored.

Purpose of the Study:

• To investigate the role of the pigeon hippocampus in model-based learning.
• To explore the neural correlates of model-based decision-making in avian brains.

Main Methods:

• Pigeons performed a two-step task with recorded hippocampal local field potentials (LFPs).
• Reinforcement learning models estimated subjective values based on behavior.
• Analysis of LFP power, dynamic correlations, and neural similarity across frequency bands.

Main Results:

• A model-based approach accurately captured pigeon behavior during learning.
• High-frequency (12-100 Hz) hippocampal power correlated with temporal context states and valuations.
• Low-frequency (1-12 Hz) hippocampal neural similarity increased for common states post-learning.

Conclusions:

• Pigeons employ model-based inferences for complex learning tasks.
• Hippocampal high-frequency oscillations support model-based valuations.
• Low-frequency neural similarity reflects learned associations between temporal states, highlighting collaborative LFP band function in avian model-based learning.